Arithmetic computation with probability words and numbers
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract Probability information is regularly communicated to experts who must fuse multiple estimates to support decision making. Such information is often communicated verbally (e.g., “likely”) rather than with precise numeric (point) values (e.g., “.75”), yet people are not taught to perform arithmetic on verbal probabilities. We hypothesized that the accuracy and logical coherence of averaging and multiplying probabilities will be poorer when individuals receive probability information in verbal rather than numerical point format. In four experiments ( N = 213, 201, 26, and 343, respectively), we manipulated probability communication format between subjects. Participants averaged and multiplied sets of four probabilities. Across experiments, arithmetic accuracy and coherence was significantly better with point than with verbal probabilities. These findings generalized between expert (intelligence analysts) and non‐expert samples and when controlling for calculator use. Experiment 4 revealed an important qualification: Whereas accuracy and coherence were better among participants presented with point probabilities than with verbal probabilities, imprecise numeric‐probability ranges (e.g., “.70 to .80”) afforded no computational advantage over verbal probabilities. Experiment 4 also revealed that the advantage of the point over the verbal format is partially mediated by strategy use. Participants presented with point estimates are more likely to use mental computation than guesswork, and mental computation was found to be associated with better accuracy. Our findings suggest that where computation is important, probability information should be communicated to end users with precise numeric probabilities.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it